In [2]:
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import pandas as pd
%matplotlib inline
import os
import pylab
import numpy as np
import umap
import glob
In [3]:
sns.set(context='paper', style='darkgrid', rc={'figure.facecolor':'white'}, font_scale=2)
In [11]:
os.chdir('/Users/s1249052/PhD/flow data/vac69a/t cells only/experiment_210618_files/csv_by_person/02')
os.getcwd()
filenames = glob.glob('*.csv')
filenames
Out[11]:
['14c01_sample1_01_0_0_T_cells.csv',
 '14c02_sample1_01_0_0_T_cells.csv',
 '14c03_sample1_01_0_0_T_cells.csv',
 '14c04_sample1_01_0_0_T_cells.csv',
 '14c05_sample1_01_0_0_T_cells.csv']
In [54]:
Baseline = pd.read_csv(filenames[0])
C_10 = pd.read_csv(filenames[1])
C_12 = pd.read_csv(filenames[2])
DoD = pd.read_csv(filenames[3])
T_6 = pd.read_csv(filenames[4])

Baseline = Baseline.apply(np.arcsinh, 'columns')
C_10 = C_10.apply(np.arcsinh, 'columns')
C_12 = C_12.apply(np.arcsinh, 'columns')
DoD = DoD.apply(np.arcsinh, 'columns')
T_6 = T_6.apply(np.arcsinh, 'columns')
In [55]:
Baseline['Timepoint'] = 'C-1'
C_10['Timepoint'] = 'C+10'
C_12['Timepoint'] = 'C+12'
DoD['Timepoint'] = 'DoD'
T_6['Timepoint'] = 'T+6'
In [56]:
concat = Baseline.append([C_10, C_12, DoD, T_6], ignore_index=True, sort=False)
In [78]:
#print(len(Baseline)+len(C_10)+len(C_12)+len(DoD)+len(T_6))
#print(len(concat))
#print(len(concat.columns))
#print(concat.columns)
#print(len(sub_concat.columns))
sub_concat.iloc[:, 1:34]
Out[78]:
141Pr_HLA-DR 142Nd_BCL-2 143Nd_CD45RA 144Nd_GZB 145Nd_CD4 146Nd_Vd2 148Nd_ICOS 149Sm_CXCR5 150Nd_CD95 151Eu_CD103 ... 168Er_CD127 169Tm_CD38 171Yb_CD49d 172Yb_CD25 173Yb_CD39 174Yb_CLA 175Lu_Perforin 198Pt_CD8 209Bi_CD16 Timepoint
0 0.000000 4.687225 3.931365 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.0 ... 4.078405 3.126917 3.486562 1.887629 1.316506 2.603347 1.417754 4.724484 0.000000 C-1
1 0.000000 3.584899 3.505189 0.896894 5.450490 2.468887 1.956988 0.000000 0.909380 0.0 ... 3.855416 5.087446 2.033088 0.870406 0.000000 0.279724 0.000000 0.697346 0.464288 C-1
2 0.159583 4.240899 0.000000 0.000000 0.574285 0.000000 0.000000 0.000000 3.566472 0.0 ... 5.064615 2.772294 3.660964 2.190205 0.953640 1.496882 4.436971 4.197494 0.000000 C-1
3 0.314058 5.029791 3.775241 0.000000 0.000000 0.000000 0.000000 0.923096 0.470137 0.0 ... 3.142208 3.463254 3.054099 1.068137 0.000000 1.207470 1.997173 5.144057 0.827061 C-1
4 0.000000 4.146923 3.147939 0.433663 0.000000 0.000000 0.000000 0.000000 0.000000 0.0 ... 3.415740 2.092585 2.641637 0.590385 1.352888 0.000000 1.348874 4.736860 0.022771 C-1

5 rows × 34 columns

In [80]:
sub_concat = concat[["In115Di","Pr141Di", "Nd142Di", "Nd143Di", "Nd144Di", "Nd145Di", "Nd146Di", "Nd148Di", "Sm149Di", "Nd150Di", "Eu151Di", "Eu153Di", "Sm154Di", "Gd155Di", "Gd156Di", "Gd158Di",
    "Tb159Di", "Gd160Di", "Dy161Di", "Dy162Di", "Dy163Di", "Dy164Di", "Ho165Di", "Er166Di", "Er167Di", "Er168Di",
    "Tm169Di", "Yb171Di", "Yb172Di", "Yb173Di", "Yb174Di", "Lu175Di", "Pt198Di", "Bi209Di", 'Timepoint']].copy()
sub_concat.columns = ['115In_CD57', '141Pr_HLA-DR', '142Nd_BCL-2', '143Nd_CD45RA', '144Nd_GZB', '145Nd_CD4', '146Nd_Vd2',
    '148Nd_ICOS', '149Sm_CXCR5', '150Nd_CD95', '151Eu_CD103', '153Eu_Va7.2', '154Sm_TIM-3', '155Gd_PD1',
    '156Gd_CD161', '158Gd_CD27', '159Tb_FoxP3', '160Gd_CTLA4', '161Dy_Tbet', '162Dy_IntegrinB7', '163Dy_CD28', '164Dy_Ki-67',
    '165Ho_CD45RO', '166Er_CD56', '167Er_CCR7', '168Er_CD127', '169Tm_CD38', '171Yb_CD49d', '172Yb_CD25', '173Yb_CD39',
    '174Yb_CLA', '175Lu_Perforin', '198Pt_CD8', '209Bi_CD16', 'Timepoint']


  
    
    

/Users/s1249052/anaconda3/lib/python3.6/site-packages/umap/spectral.py:229: UserWarning: Embedding 2 connected components using meta-embedding (experimental)
  n_components

CPU times: user 4min 31s, sys: 11.9 s, total: 4min 43s
Wall time: 4min 1s
In [84]:
reducer = umap.UMAP(n_neighbors=20, min_dist=0.1, n_components=2, metric='euclidean')
%time embedding = reducer.fit_transform(sub_concat.iloc[:, 1:34])
embedding.shape 
sub_concat['umap1'] = embedding[:, 0]
sub_concat['umap2'] = embedding[:,1]

channels=list(sub_concat.columns.values)

/Users/s1249052/anaconda3/lib/python3.6/site-packages/umap/spectral.py:229: UserWarning: Embedding 2 connected components using meta-embedding (experimental)
  n_components

CPU times: user 3min 48s, sys: 11.9 s, total: 4min
Wall time: 3min 20s
In [91]:
channels=list(sub_concat.columns.values)

for n in channels[0:34]:
    cmap=plt.get_cmap('nipy_spectral')
    fig=plt.figure(figsize=(15,10))
    ax=plt.subplot()
    ax.scatter(sub_concat['umap1'], sub_concat['umap2'],c=sub_concat[n], cmap=cmap, s=1)
    plt.title('{}'.format(n))
    #plt.savefig('{}.pdf'.format(n))
    plt.show()
In [101]:
Baseline_csv = sub_concat[sub_concat['Timepoint'] == 'C-1']
C_10_csv = sub_concat[sub_concat['Timepoint'] == 'C+10']
C_12_csv = sub_concat[sub_concat['Timepoint'] == 'C+12']
DoD_csv = sub_concat[sub_concat['Timepoint'] == 'DoD']
T_6_csv = sub_concat[sub_concat['Timepoint'] == 'T+6']
In [ ]:
Baseline_csv.to_csv('baseline_02_UMAP.csv', sep=',') 
C_10_csv.to_csv('C+10_02_UMAP.csv', sep=',') 
C_12_csv.to_csv('C+12_02_UMAP.csv', sep=',') 
DoD_csv.to_csv('DoD_02_UMAP.csv', sep=',') 
T_6_csv.to_csv('T+6_02_UMAP.csv', sep=',')
In [ ]:
cytobank.to_csv('timecourse_02.csv', sep=',')
In [24]:
#cytobank=flow_data[["Cd114Di", "In115Di","Pr141Di", "Nd142Di", "Nd143Di", "Nd144Di", "Nd145Di", "Nd146Di", "Sm147Di", "Nd148Di", "Sm149Di", "Nd150Di", "Eu151Di", "Eu153Di", "Sm154Di", "Gd155Di", "Gd156Di", "Gd158Di",
"Tb159Di", "Gd160Di", "Dy161Di", "Dy162Di", "Dy163Di", "Dy164Di", "Ho165Di", "Er166Di", "Er167Di", "Er168Di",
"Tm169Di", "Yb171Di", "Yb172Di", "Yb173Di", "Yb174Di", "Lu175Di", "Yb176Di", "Pt198Di", "Bi209Di"]].copy()

#cytobank.columns = ['114Cd_CD14', '115In_CD57', '141Pr_HLA-DR', '142Nd_BCL-2', '143Nd_CD45RA', '144Nd_GZB', '145Nd_CD4', '146Nd_Vd2',
'147Sm_CD20', '148Nd_ICOS', '149Sm_CXCR5', '150Nd_CD95', '151Eu_CD103', '153Eu_Va7.2', '154Sm_TIM-3', '155Gd_PD1',
'156Gd_CD161', '158Gd_CD27', '159Tb_FoxP3', '160Gd_CTLA4', '161Dy_Tbet', '162Dy_IntegrinB7', '163Dy_CD28', '164Dy_Ki-67',
'165Ho_CD45RO', '166Er_CD56', '167Er_CCR7', '168Er_CD127', '169Tm_CD38', '171Yb_CD49d', '172Yb_CD25', '173Yb_CD39',
'174Yb_CLA', '175Lu_Perforin', '176Yb_CX3CR1', '198Pt_CD8', '209Bi_CD16']
In [38]:
#copy with the channels that worked/are interesting

cytobank=flow_data[["In115Di","Pr141Di", "Nd142Di", "Nd143Di", "Nd144Di", "Nd145Di", "Nd146Di", "Nd148Di", "Sm149Di", "Nd150Di", "Eu151Di", "Eu153Di", "Sm154Di", "Gd155Di", "Gd156Di", "Gd158Di",
"Tb159Di", "Gd160Di", "Dy161Di", "Dy162Di", "Dy163Di", "Dy164Di", "Ho165Di", "Er166Di", "Er167Di", "Er168Di",
"Tm169Di", "Yb171Di", "Yb172Di", "Yb173Di", "Yb174Di", "Lu175Di", "Pt198Di", "Bi209Di"]].copy()

cytobank.columns = ['115In_CD57', '141Pr_HLA-DR', '142Nd_BCL-2', '143Nd_CD45RA', '144Nd_GZB', '145Nd_CD4', '146Nd_Vd2',
'148Nd_ICOS', '149Sm_CXCR5', '150Nd_CD95', '151Eu_CD103', '153Eu_Va7.2', '154Sm_TIM-3', '155Gd_PD1',
'156Gd_CD161', '158Gd_CD27', '159Tb_FoxP3', '160Gd_CTLA4', '161Dy_Tbet', '162Dy_IntegrinB7', '163Dy_CD28', '164Dy_Ki-67',
'165Ho_CD45RO', '166Er_CD56', '167Er_CCR7', '168Er_CD127', '169Tm_CD38', '171Yb_CD49d', '172Yb_CD25', '173Yb_CD39',
'174Yb_CLA', '175Lu_Perforin', '198Pt_CD8', '209Bi_CD16']
In [39]:
len(cytobank.columns)
Out[39]:
34
In [41]:
cytobank=cytobank.apply(np.arcsinh, 'columns')

reducer = umap.UMAP(n_neighbors=30, min_dist=0.01, n_components=2, metric='euclidean')

%time embedding = reducer.fit_transform(cytobank)
embedding.shape  

CPU times: user 41.1 s, sys: 2.42 s, total: 43.5 s
Wall time: 41 s
Out[41]:
(27003, 2)
In [42]:
cytobank['umap1'] = embedding[:, 0]
cytobank['umap2'] = embedding[:,1]
In [43]:
channels=list(cytobank.columns.values)

for n in channels:
    cmap=plt.get_cmap('nipy_spectral')
    fig=plt.figure(figsize=(15,10))
    ax=plt.subplot()
    ax.scatter(cytobank['umap1'], cytobank['umap2'],c=cytobank[n], cmap=cmap, s=1)
    plt.title('{}'.format(n))
    #plt.savefig('{}.pdf'.format(n))
    plt.show()
In [21]:
os.getcwd()
Out[21]:
'/Users/s1249052/PhD/flow data/vac69a/t cells only/experiment_210618_files'
In [44]:
cytobank.to_csv('vac69a_03_t+6_umap.csv', sep=',')